<![CDATA[Uneven participation in employment-based social protection remains a major challenge in decentralized labor markets where regional economic structures, labor informality, and institutional capacity vary substantially across locations. Conventional panel regression and standard spatial econometric models generally assume homogeneous relationships across regions, potentially obscuring localized determinants of participation behavior. This study examines spatially varying determinants of BPJS Ketenagakerjaan participation in South Kalimantan Province, Indonesia, using Geographically Weighted Panel Regression (GWPR) applied to panel data from thirteen districts and cities during 2018–2022. The GWPR approach is employed because it allows regression coefficients to vary across space and time, enabling the identification of spatial nonstationarity that cannot be captured by global panel models. The results reveal clear spatial heterogeneity in participation dynamics. The number of registered companies emerges as the most consistent determinant, showing statistically significant positive effects across all districts with coefficients ranging from 0.099 to 0.143. In contrast, informal worker income demonstrates localized negative effects in several districts (−0.104 to −0.088), suggesting substitution between informal earnings and participation in formal protection schemes. Average years of schooling shows strong positive effects in selected regions (0.554–0.699), indicating the importance of human capital in increasing social insurance awareness and participation. Model adequacy testing further confirms that the GWPR specification provides a better representation of spatial variation than the global panel model.]]>
